The investigators will extend their laboratory study of the effects of off-fault damage on earthquake rupture propagation from 6x6 inch birefringent polymer plates to 12x12 inch marble plates. For marble, the Rice et al. (2005) model predicts that the interaction of the crack tip with initial damage should extend to distances on the order of 10 cm from the fault plane, in contrast to an interaction distance of about 1 cm in the polymer. The main advantage of using marble is that it has a uniform distribution of 1 mm grains with weak boundaries that can be exploited to produce a uniform millimeter-scale array of initial damage. This combination in marble of a larger damage zone with smaller scale initial damage should allow for the observation of the zone close to the crack tip where models predicts that new damage and granulation should occur.
The larger marble plates and nucleation from one edge are also required to keep the rupture from nucleating an out-of-plane tensile wing crack. They have already constructed a larger loading frame and successfully nucleated and propagated mode II shear ruptures on the fault plane in these larger plates. Larger plates are also required to study the larger damage band predicted for marble.
The data will be interpreted using the dynamic slip pulse model recently developed by Rice et al. (2005). In this model slip is limited to a finite length of the fault plane, which has the effect of limiting the width of the zone of off-fault damage to a finite value as the rupture speed approaches the limiting Rayleigh speed. The micromechanical model of Ashby and Sammis (1990) will allow the researchers to distinguish between Coulomb slip on preexisting fractures and the generation of new damage, which leads to local failure and the formation of gouge and breccia. The researchers are currently collaborating with Jim Rice and Elizabeth Templeton to cast the Ashby-Sammis damage mechanics into a form that is synergistic with the elastic-perfectly plastic model. The ultimate goal, to model a freely propagating slip pulse in a realistic non-linear medium, remains an unsolved problem. The laboratory experiments proposed here will serve to guide the development of such a model.
